In February 2026, the FDA made two announcements that will fundamentally change how drugs are developed, approved, and brought to market.  

The first was a new guidance on individualized therapies for ultra-rare genetic diseases, establishing real-world evidence as the expected evidentiary pathway for approval. The second: a New England Journal of Medicine article by the head of the FDA declaring that one pivotal trial is now the default standard for drug approval, ending what the agency itself called “the two-trial dogma.” 

Together, these represent a sea change in the use of real-world data in drug and device approval. This is revolution, not evolution. 

How We Got Here 

Many people think the use of real-world evidence in a regulatory context is novel. The truth is, the FDA has been using it for decades. Drugs have had their approval revoked based on real-world safety signals. And the FDA has accepted RWE to demonstrate efficacy, but largely as a one-off when a traditional clinical trial wasn’t feasible or when there was significant unmet or public health need. Led by Dr. Richard Pazdur, the oncology division has been at the forefront of the regulatory use of real-world data, closely followed pediatric and rare disease. 

What has changed is the sophistication of the science. The design and analysis of real-world evidence has matured and the FDA has responded to those advances by accepting real-world evidence to a greater degree. At the same time, advances in molecular genetics and cell technologies have given the agency more confidence in the mechanistic understanding of how drugs work. These two forces, better real-world evidence and deeper biological understanding, converged to make this moment possible. 

There is also a practical dimension. The FDA cited “lowering capital costs for drug developers” as an important reason for this change. That language would have been highly unusual from the FDA at any point before this.

What Was Said: The Individualized Medicine Guidance

The FDA’s new Plausible Mechanism Framework guidance addresses individualized therapies for genetic conditions with a known biological cause. These are ultra-rare diseases where the patient population may be a handful of people and traditional randomized controlled trials are not feasible. 

The guidance establishes that for these conditions, real-world evidence can be seen as the standard for approval. External controls and natural history studies are now the expected approach. The FDA is directing companies toward this pathway, not merely accommodating it on a case-by-case basis. 

This is a first. Randomized, double-blind placebo-controlled have always been the default.  For regulatory and clinical operations professionals who have spent their careers running traditional trials, that is a significant shift and raises a plethora of questions: how do you ensure equivalence at baseline if you don’t randomize? What does fit-for-purpose mean? How do you translate real-world data into MedDRA and CDISC? 

What Was Said: One Trial as the New Default 

The standard for the FDA since 1962 wasn’t just a randomized, double-blind placebo- controlled trial but two such trials. Two adequate and well-controlled trials. The NEJM article, authored by the Commissioner and Chief Medical/Scientific Officer, declared that one pivotal trial is now the default for approval. The article names real-world evidence specifically as qualifying confirmatory evidence, alongside mechanistic science, animal models, and data from related indications. 

In an unusual departure, the Commissioner stated “lowering capital costs for drug developers” was an important consideration for this new paradigm, and that “Our move to change the FDA’s default position from two clinical trials to one will substantially reduce costs for sponsors and will speed drugs to market.” The economics are straightforward. A single pivotal trial costs between $30 million and $150 million, not including the costs of creating and filing the NDA/BLA. A fully real-world confirmatory study can be executed for 15% to 20% of that cost, including filing. The basic math will drive adoption. 

Another unusual aspect of the NEJM announcement is that there was no accompanying regulation or draft guidance published in the Federal Register proposing how this change would be implemented or seeking scientific and community feedback. We have been inundated with requests from companies to speak to them about what the FDA’s expectations are concerning the implementation of these changes. 

The good news is, the FDA has released 9 guidances on how real-world data can and should be used in a regulatory setting. But as with all such guidances, the key is understanding how to implement them. In the words of another FDA commissioner regarding these guidances: “Your job isn’t to find the holes in them, your job is to help the agency fix those holes.” 

What It Means 

Taken together, these two announcements create a new reality for drug development. Companies will do one trial, bring their drug to market, and then generate real-world confirmatory evidence while patients are already benefiting from the therapy and the company is generating revenue to fund the work. For small companies that have been surviving on venture capital, that changes the economics of development entirely. 

Here is what a real-world confirmatory study looks like in practice. You could conduct a second, randomized, double-blind placebo-controlled trial supplementing patients randomized to placebo with real-world external/synthetic controls. You could conduct a single arm study with fully external/synthetic controls. Or you could conduct a fully real-world trial emulation. The last is the fastest and most cost-effective. 

Regardless of the study, the very first step is to demonstrate that the proposed real-world data are fit-for-purpose, or more recently the FDA has referred to this as fit-for-use. There is no such thing as a dataset being generally fit-for-use, it is specific to the indication and to the drug. In addition to the data being relevant to the use case, data provenance, traceability, transformation, verification and quality must all be assessed. A first Type C meeting should be prepared with a proposed study design and a fit-for-purpose validation package for the Agency to review. 

If the FDA agrees, then you start the process of creating a protocol. This should be done withing whatever CRM system you use for clinical trials and in which you’ll maintain the trial master file, most often Veeva Vault. Like a clinical trial, traceability is key.  

The first question trial biostatisticians ask us is: how do you ensure baseline equivalence if you don’t randomize? With real-world data, we use propensity scores to generate inverse probability of treatment or standardized morbidity-mortality weighting. Here, we collaborate with a company’s real-world evidence team to build a bridge with biostats to design a statistical analysis plan incorporating the skills and expertise from both XX of the organization. 

From the real-world evidence teams, we’re asked: Why can’t we run this? Here we build a bridge to the FDA regulatory requirements for data management and analysis. 

And we work with both teams to set up the required infrastructure for a Clean Room Committee, to understand how diagnosis codes are translated into MedDRA for safety analyses, and how all the real-world data are translated into CDISC for the production of final TLFs. We work to address the technical challenges that come with validating real-world datasets with Pinnacle 21, and eCTD publishing with Veeva RIM, and submission through the FDA portal. 

What Companies Should Be Thinking About 

The most common reaction we hear from regulatory and clinical operations teams is that real-world evidence approvals are still the exception. That the FDA isn’t really ready for this. That it’s too risky to build a development program around it. 

These announcements make that position untenable. The FDA has told the industry directly: this is the expected approach for individualized therapies, and it is the default for confirmatory evidence across drug development more broadly. Companies that continue to operate under the old assumptions will watch their competitors get to market faster and cheaper. 

But the enthusiasm has to be tempered by the reality of execution. The FDA expects real-world evidence studies to be treated with the same rigor as traditional clinical trials. There are now eight or nine guidances that lay out exactly how. Companies must prespecify endpoints, register the study, submit a protocol through the same review and approval process. The agency expects data provenance, traceability, transparency, and auditability. They expect companies to demonstrate that the data is fit for purpose for the specific disease and dataset. 

This is where most companies run into trouble. The skillsets required to do this work are split across two groups that typically do not work together. Real-world evidence scientists understand the epidemiology, the propensity score methods, the claims database structures. Clinical development teams understand the regulatory environment, the SOPs, the FDA’s expectations for how data must be managed and submitted. But those two worlds operate with very different processes, vocabularies, and standards. 

For smaller companies, one or both of those capabilities may not exist internally at all. For mid-sized companies with internal RWE teams, those teams often have no experience working in a regulatory environment. And for clinical trial teams at companies of any size, real-world evidence methods are frequently unfamiliar territory. 

There is also a technology dimension that is easy to underestimate. The FDA’s own systems are still adapting to accommodate real-world datasets at scale. In our experience, submitting real-world evidence to the FDA has been a learning process for the agency itself. The guidances lay out what the FDA wants, but there are practical gaps between theory and implementation. The FDA has been direct about this: it is the submitting company’s job to solve those problems, not to point them out. 

Where Slipstream Fits 

Slipstream’s Digital CRO practice was built before these announcements, specifically for the kind of evidence generation the FDA is now making standard. 

We bridge the gap between real-world evidence science and clinical trial regulatory standards. We bring regulatory SOPs to RWE teams and introduce real-world design and analytic methods to clinical trial teams. For companies without internal RWE capability, we provide it entirely. 

Our team includes epidemiologists, data scientists, biostatisticians, programmers, medical writers, and regulatory experts. We operate within compliant systems where every change is tracked and auditable, with more than 250 SOPs that guide how data is managed from intake through FDA submission. Our partnership with Komodo Health gives our clients access to a 330-million-patient healthcare claims database that is transferable to the FDA while maintaining HIPAA compliance. 

We have led FDA submissions with real-world evidence as the primary basis for approval. We understand the practical challenges the guidances do not fully address. And we operate as both a technology company and a life sciences company, which matters because the intersection of those two capabilities is exactly where the hardest problems in this space live. 

The FDA just formalized what we have been building toward. The question for companies now is whether they have the right partner to execute.